Urbanization in Chinain the1990s: Patterns and Regional Variations
Zai Liang
Department of Sociology
StateUniversity of New York at Albany
Hy Van Luong
Department of Anthropology
University of Toronto
and
Yiu Por Chen
DePaulUniversity
*Paper prepared for conference sponsored by Urban China Research Network of SUNY Albany, New Orleans, January 15-16, 2005. This research is supported, in part, by a FIRST Award from the National Institute of Child Health and Human Development (1R29HD34878-01A2). We are grateful to Grace Kao, Michael J. White, and Tufuku Zuberi for their comments on earlier versions of the paper.
Introduction
Demographically speaking, the 1980s and the 1990s were the best of times for migrants in China. After several decades of control and retrenchment, Chinese cities are experiencing an unprecedented increase in their migrant populations. From the most desirable migrant destinations of Beijing, Shanghai, and Guangdong to remote areas of Xinjiang and Heilongjiang, migrants are everywhere and changing the face of Chinese urban landscape. In response to the large volume of migrant population, students of migration in China have been studying different aspects of this great wave of migration (Goldstein, Goldstein, and Guo; 1991; Hare, 1999; Liang and Ma, 2004; Roberts, 1997; Yang, 1994). Since a large volume of China’s recent migrant population is characterized by rural to urban migration, migration clearly has consequences for China’s urbanization patterns. Indeed, in the words of Rogers and Williamson (1984), “urbanization is attributable to mobility revolution, the transformation experienced by societies with low migration rates as they advance to a condition of high migration rates (p.262).”
Similar to the study of migration, for a long time, research on urbanization in China was hampered by the lack of systematic and adequate data, until the late 1970s with the advent of economic reforms. With the release of results from China’s 1982 and 1990 Population Censuses (recently the 2000 census), along with other data on China urban statistics, studies of Chinese urbanization has significantly increased. Previous studies of Chinese urbanization cover a wide-range of issues and clarified much confusion regarding data and definitions of Chinese urbanization and provided many benchmarks from which future researchers can draw on. Chan (1994), Ge (1999), and Kirby (1985) published in-depth historical overview of urbanization patterns and dynamics dating back to 1949 (Chan, 1994; Ge, 1999; Kirby, 1985). Goldstein (1985) detailed analysis of characteristics of rural and urban population and level and patterns of urbanization. Other important scholar work includes analysis of regional patterns of urbanization focusing on the Pearl River Delta region (Guldin, 1992; Wang, 1996; Xu, 1990; Ye, 1990); a demographic accounting of China’s urban growth (Chan, 1994; Chan and Hu, 2003; Chen et al., 1998); as well as strategic discussions and planning of China future roads of urbanization (Ge, 1999; Wang and Sun, 1997).
In this chapter, capitalizing on the recent data from the 1990s, we provide an overview of regional patterns of urbanization and focus on the demographic sources of urbanization in China. The central question we address is the extent to whichChina’s recent urbanization can be explained by migration/boundary reclassification vs. urban natural increase. Our study is expected to contribute to the current literature on urbanization in China in several aspects. First, in some of the most influential international studies of demographic sources of urbanization, China is often excluded because of lack of data. For example, in Preston’s 1979 study, China was not included (Preston, 1979). This changed as more data from China became available. The recent study by Chen et al. (1998) used data from the 1982 and 1990 China Population Censuses, but not after. Our study will update Chen et al.’s (1998) work by using data from the 1990s.
Second, most studies of demographic sources of urban growth by and large use country as unit of analysis. This makes sense if a country is relatively homogeneous. However, given the large regional variation in socio-economic development, migration, and fertility patterns in China, we plan to carry out this exercise one step further, i.e. to examine urbanization patterns at the province level. As Zhou and Ma (2003) suggested that data from different provinces may not be comparable because provinces may have different ways of adopting urban administrative changes. If so, it would give more justification for conducting this exercise at the province level.
Third, the current exercise also has theoretical significance because of its potential contribution to theorizing the experience of urbanization in different parts of the world, especially the developing world. For example, recent work by James Lee and Wang Feng (2001) challenges the traditional Malthusian notion of Chinese demographic behaviors (universal marriage, high fertility and high mortality), which has been taken for granted until recently. By the same token, Preston’s 1979 article has been widely cited as accepted wisdom of describing urbanization experience of developing countries (also see Brokerhoff, 2000). However, the article does not include China for lack of data. Furthermore, the article was written 25 years ago during which time fertility has dropped significantly in many parts of the world especially in Asian countries. It is high time that we provided some update with new empirical evidence.
Finally, in light of this new thinking on patterns of urbanization in less developed countries, we will use some recent data from Southeast Asia and especially Vietnam to demonstrate migration-dominated pattern of urbanization is by no means unique to China, it is happening in other parts of Asian countries as well.
Regional Patterns of Urbanization in China, 1990 and 1995
Unlike many other developing countries, there are tremendous regional variations in China. To begin with, China’s population is unevenly distributed, which has been noted by many scholars (Goldstein 1985; Hu, 1982). Aside from population, there is also important variation in socioeconomic development. With transition to a market-oriented economy, regions with easy access to market and capital are able to progress whereas the other regions either remain static or fall further behind. The well-known story is the inequality between coastal and non-coastal regions. This inter-regional inequality has clear implications for urbanization. One major pattern of migration that characterizes the 1990s is the large concentration of migrants in the coastal regions, especially coastal cities. For example, since 1987, migration to the coastal region has accelerated, especially to cities. Thus, we expect that the urbanization in the coastal region is likely to experience a more rapid growth than other regions.
We take two steps to examine regional variations in urbanization. First, we describe urbanization level in 1995 by province and examine changes from 1990 to 1995. Second, we will examine variations of impact of migration on urbanization in next section. For China as a whole, urbanization level had an increase of a modest 2.34 percent over the period between 1990 and 1995. The annual rate of growth during this time was 1.72 percent, which is consistent with corresponding rate of 1.7 for other developing countries during 1975-1995 (Chen et al., 1998).[1] However, there is tremendous regional variation in the pace of urbanization. The biggest increase in the level of urbanization is found in Shanghai, from 66.14 percent in 1990 to 83.87 percent in 1995, an increase of 17.6 percent! [2] We suggest that this dramatic increase in urbanization in Shanghai is largely driven by Shanghai’s recent development of “Pudong” (across the HuangpuRiver), an enterprise zone mainly developed by large transnational companies. Today’s Pudong looks like New York’s Manhattan on a smaller scale. A 1993 Survey of Floating Population in Shanghai found that more than one third of Shanghai’s floating population work in development zones, including Pudong district (Wang, 1995). Migrant workers are employed in many occupations related to Pudong’s development, including construction, transnational corporations, and service related jobs. Knowing that the coastal region has attracted so many migrants, we now look at changes in urbanization between the coastal region and non-coastal region. [3] For example, mean annual growth rate for the coastal region as a whole was 1.54 percent while mean annual growth rate for non-coastal region was 1.27 percent. However, this coastal vs. non-coastal difference is actually dominated by Shanghai, whose urbanization level increased dramatically in the first part of the 1990s. If we exclude Shanghai from the coastal region, the mean annual growth rate would be only 1.24 percent, even slightly lower than that of non-coastal regions.[4] This suggests two things: first is that even within coastal provinces, there is much heterogeneity, which makes generalizations very difficult; secondly, with a few exceptions, migrant population still constitutes only a small proportion of a province’s population and most of the migrants went to cities. It should be noted that urbanization level is not only determined by the population living in the urban areas, but is also determined by the size of rural population in a locale.
Also of interest is the case of Guangdong, located in the Pearl Delta region and most likely due to its close connections with Hong Kong, Guangdong has experienced a very solid economic growth during the economic reform era. It is perhaps the most attractive place for migrants, receiving the largest migrant population volume in China. Therefore, it may surprise some readers to find that urbanization levels in Guangdong only increased slightly from 38.92 percent in 1990 to 39.28 percent in 1995. We argue that this is largely attributable to Guangdong’s relatively high level of fertility, especially in rural areas. In 1990, for example, Guangdong’s crude birth rates in rural and urban areas were 24.99 and 17.98 respectively. The corresponding rates for China as a whole in 1990 were 22.73 and 16.73.[5] Our later analyses of the source of urban population growth indicates that natural increase explains more than half of the urban increase between 1990 and 1995.
One may also be puzzled by the fact that Beijing’s level of urbanization actually declined slightly from 73.22 percent in 1990 to 69.75 percent in 1995. Since floating population accounts for 15 percent of Beijing’s total population in 1995, we suggest this decline in urbanization level is due in large part to the residential distribution of the floating population. Our calculations (not shown here) reveal that in 1990 14.7 percent Beijing’s floating population lived in rural areas. In contrast, by 1995, the corresponding percentage nearly doubled to 29 percent. With more and more migrants moving to Beijing, the housing is saturated which has forced many migrants to move to nearby rural areas. Thus, while holding on to jobs in Beijing, migrants can enjoy the low cost of living in the rural areas.
While many researchers have focused on fast-developing coastal regions, our results suggest other non-coastal provinces also experienced a solid increase in the level of urbanization during this time period. The following non-coastal provinces saw their level of urbanization increased by more than 5 percent: Hunan, Sichuan, Shannxi, and Jilin. With the exception of Jilin province, the other three provinces started with a relatively low level of urbanization in 1990.
In sum, then, the 1990s were accompanied by a modest gain in the level of urbanization in China. Whereas coastal regions exhibit variations in urbanization, some non-coastal provinces experienced rather solid gains. It should be noted that urbanization level is not only affected by migration (as in the case of Beijing) but also affected by the dynamics of fertility and mortality, as well as major economic development initiative (as in the case of Shanghai).
Migration, Fertility, and Demographic Sources of Urbanization
The text-book version of migration and urbanization can roughly be restated in the following way. The standard economic analysis of migration and development suggests the following reasoning (Lewis, 1954 and Todaro, 1969). The Industrial Revolution introduced new technological innovations, which increased agricultural productivity. The technological innovations, along with the enclosure movement lowered the demand for farm labor and made peasants increasingly redundant in agricultural production, therefore peasants migrated to cities in order to obtain manufacturing jobs and higher wages. According to these views, moreover, the excess labor supply is eventually absorbed into the newly created jobs. In this process income differential between rural and urban areas is reduced which in turn causes migration to reach equilibrium (Williamson, 1988; Portes and Benton, 1984; Rogers and Williamson, 1984).
The urbanization experience of European countries suggested that in most cases migration was the major contributing factor to urbanization. Using data from the Great Britain during the period of 1776 to 1871, Williamson (1988) provided empirical support for this argument. In the years during initial stage of the Industrial Revolution, rural to urban migration accounted for more than 50% of urban growth. In fact for the periods of 1781-1786 and 1801-1806, rural to urban migration accounted for nearly 90% of the urban growth. For most of remaining years during 1776-1871 (roughly a century), rural to urban migration contributed about 50% to urban growth. Thus Williamson (1988) concluded that “…immigration was a much more important source of city growth during the First Industrial Revolution than it is in the contemporary Third World (p.293).”
Underlying this pattern of urban growth is the notion that urbanization in developed countries historically happened at a time when fertility in urban areas experienced a major decline, therefore rural to urban migration became a leading factor in urban growth. In contrast, it is argued that in today’s developing countries, urbanization is taking place at a time when fertility level in urban areas remains quite high. Therefore the demographic sources of urbanization in less developing countries are different from that of more developed countries. As early as the mid-1960s, Kinsley Davis (1965) argued that urban natural growth contributed much more than rural-urban migration to the level of urbanization in the developing countries. Research by Preston (1979) provided a more systematic empirical support for this generalization. Preston (1979) wrote: “of the 29 developing countries whose data support a decomposition of the sources of urban growth during the most intercensal periods, 24 had faster rates of urban natural increase than of net in-migration (p.198).” Thus, it is clear that the major source of urbanization in most of today’s less developed countries is driven by natural increase rather than migration.
Urban fertility is certainly high in some of the African countries (citation?), the story for China in the 1990s is clearly different. We argue that although China is still a developing country by economic measures, it may not follow the patterns of urbanization of other developing countries for two main reasons. First, China’s family planning program and policies have played a very important role in fertility decline both in urban and rural areas. Since the 1970s China has adopted a series family planning policies and the most well-know one being “the one child policy” which began in 1979 (Banister, 1986). In the minds of policy-makers, these policies are necessary for the well-being of the Chinese population and for the smooth progress of China’s economic development program (Lee and Wang, 1999). To implement the one child policy, a reward and penalty system is also designed to reward individuals who pledge having only one child and to punish couples who decide to have more than one child. The penalty for non-compliance among cadres and family planning worker is especially severe (Hardee-Cleveland and Banister, 1988). As expected, China’s “One Child Policy” has had a tremendous impact on fertility decline in China, more so in urban areas than rural areas. Evidence suggests that this policy has been much more successful in cities than in the countryside (Retherford et al., 2004). As a result, the established TFRs are as follows: 1.11-1.19 for cities, 1.47 for towns, and 1.84-2.00 for rural areas for the period of 1990-1995 (Retherford, 2005, p.25). Aside from one child policy, China has one of the highest contraceptive use in the world (Lee and Wang, 1999)—among married women 90% of them use contraceptives in 1990. Therefore we argue that fertility in urban China in the 1990s is so low that its contribution to urban growth is quite small, especially compared to the impact of migration.
Second, since the early 1980s, there has been a major increase in China’s migrant population, especially the so called “floating population”, i.e. migrants who do not have household registration status in places of destination. Data from censuses and survey show that the floating population (cross-county) has increased from about 22 million in 1990 and by 1995 it has more than doubled. This increase in migration is driven by several factors. One is the implementation of “household responsibility system” in the countryside that has greatly increased the efficiency of agricultural production and created a large number of surplus labor in the countryside. Perhaps more importantly, China’s transition to a market economy has led to major changes in the way in which cities are operating, so much so that peasants no longer need to obtain urban household registration status to stay and work in cities. At the same time, booming joint enterprises, private businesses, service industries especially along the booming coastal regions create constant demand for labor (Liang, 2001; Liang and Ma, 2004). As a result, any visitors to China today cannot fail to see migrants in cities. Recent empirical data do suggest this intuition. To take two extreme examples, floating population in Shanghai and Guangdong account for about a quarter of their respective population in 2000. In sum, the combination of low fertility and large volume of migrant population in the 1990s prompt us to hypothesize that urban population growth in China in the early 1990s is mainly driven by migration rather than natural increase.